Upcoming Events
Vendor-Neutral GPU Programming in Chapel
July 31, 2024
Writing programs on modern computers requires parallelism to achieve maximum performance. This is complicated by GPUs, which provide great parallel performance at the price of more complex programming. Chapel is an open-source parallel programming language that supports portable, performant software on CPUs and GPUs using a single unified set of language features. In this talk, we will showcase Chapel's vendor-neutral GPU support and share user experiences writing GPU-enabled programs in Chapel and show how you can write vendor-neutral GPU programs today. How you can get involved in our open-source work and our future plans will conclude this talk.
LLM finetuning for mere mortals
August 28, 2024
Everyone wants to use Large Language Models (LLMs), and for good reason. With applications ranging from content creation to automated software development, LLMs have the potential to transform nearly every industry. How can you make the most of this technology when applying it to your own use-cases? Finetuning is one highly effective approach, but can be challenging to implement correctly. In this talk, you'll learn about the challenges involved, and how us mere mortals can tackle them using HPE's new software that leverages the open-source machine learning (ML) ecosystem.
Enhancing NLP with Retrieval-Augmented Generation: A Practical Demonstration
September 18, 2024
In the evolving landscape of Natural Language Processing (NLP), Retrieval-Augmented Generation (RAG) stands out as a powerful technique that enhances the NLP application by incorporating relevant external information. In this session, we will delve into the fundamentals and applications of RAG, providing a comprehensive overview of how it integrates retrieval mechanisms with generative capabilities to produce more accurate and contextually aware responses. We will then transition into a live demonstration showcasing the practical usage of RAG and the process of augmenting a generative model with external knowledge sources, showcasing how RAG improves the relevance and quality of generated outputs in real-time applications.